JOURNAL OF MORPHOLOGY 267:1338–1355 (2006)

The Laterophysic Connection and Swim Bladder of Butterflyfishes in the (: Chaetodontidae)

Jacqueline F. Webb,1* W. Leo Smith,1–3 and Darlene R. Ketten4,5

1Department of Biology, Villanova University, Villanova, Pennsylvania 19085 2Department of , American Museum of Natural History, New York, New York 10024 3Center for Environmental Research and Conservation, Columbia University, New York, New York 10027 4Department of Biology, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543 5Department of Otology and Laryngology, Harvard Medical School, Boston, Massachusetts

ABSTRACT The laterophysic connection (LC) is an standing of how fishes interpret sound. An acoustic association between bilaterally paired, anterior swim stimulus (e.g., modeled as a vibrating sphere, Kal- bladder extensions (horns) and medial openings in the mijn, 1989) has two components—hydrodynamic supracleithral lateral line canals that diagnoses butterfly- flow (the ‘‘near field’’) and a propagating sound pres- fishes in the genus Chaetodon. It has been hypothesized sure wave (the ‘‘far field’’). Hydrodynamic flow is that the LC makes the lateral line system sensitive to sound pressure stimuli that are transmitted by the swim generated by the movement of water near the bladder horns and converted to fluid flow into the lateral acoustic stimulus source and sound pressure waves line system via a laterophysic tympanum. The purpose of propagate from the acoustic source as a cyclic com- this study was to define variation in the morphology of pression and rarefaction of water molecules. The the LC, swim bladder and swim bladder horns among 41 mechanosensory lateral line is generally sensitive Chaetodon from all 11 Chaetodon subgenera and a to hydrodynamic flow (local displacement) within 1– species from each of four non-Chaetodon genera using 2 body lengths from the source. The inner ear is also gross dissection, histological analysis as well as 2D or 3D sensitive to hydrodynamic flow as a result of whole CT (computed tomographic) imaging of live, anesthetized body acceleration, but sound pressure-induced oscil- fishes. Our results demonstrate that the lateral line sys- lations of the air volume within the swim bladder tem appears rather unspecialized with well-ossified nar- row canals in all species examined. Two LC types (direct generates a secondary local displacement field that and indirect), defined by whether or not the paired ante- is capable of stimulating the inner ear (reviewed by rior swim bladder horns are in direct contact with a Schellart and Popper, 1992; Popper and Fay, 1999; medial opening in the supracleithral lateral line canal, Popper et al., 2003). are found among species examined. Two variants on a The ability to detect sound pressure stimuli is direct LC and four variants of an indirect LC are defined enhanced by the presence of intimate associations by combinations of soft tissue anatomy (horn length [long/ of a volume of air (in the swim bladder, swim blad- short] and width [wide/narrow], number of swim bladder der horns or branchial diverticulae, reviewed by chambers [one/two], and presence/absence of mucoid con- Schellart and Popper, 1992) and the otic capsule, nective tissue in the medial opening in the supracleith- known as otophysic connections, which increase the rum). The combination of features defining each LC vari- ant is predicted to have functional consequences for the auditory sensitivity and frequency response of the bioacoustics of the system. These findings are consistent inner ear (Poggendorf, 1952; Coombs and Popper, with the recent discovery that Chaetodon produce sounds 1979) and the distance over which sound pressure during social interactions. The data presented here pro- stimuli can be detected (Coombs et al., 1992, Popper vide the comparative morphological context for the and Fay, 1993, Coombs and Montgomery, 1999). functional analysis of this novel swim bladder-lateral Otophysic connections occur in at least 50 fami- line connection. J. Morphol. 267:1338–1355, 2006. Ó 2006 lies in all major teleostean lineages (Schellart and Wiley-Liss, Inc.

KEY WORDS: Chaetodon; lateral line; hearing; swim Contract grant sponsor: National Science Foundation; Contract bladder; laterophysic connection; CT imaging; butterflyfish grant numbers: IBN-9603896, IBN-0132607.

*Correspondence to: Jacqueline F. Webb, Department of Biological Sound stimuli are important in the social behav- Sciences, University of Rhode Island, 100 Flagg Road, Kingston, RI 02881. E-mail: [email protected] ior of a wide variety of fishes (Myrberg, 1981; Haw- kins, 1993; Ladich and Bass, 2003; Ladich and Pop- Published online 18 October 2006 in per, 2004), but the complex physical features of Wiley InterScience (www.interscience.wiley.com) underwater acoustics have confounded our under- DOI: 10.1002/jmor.10480

Ó 2006 WILEY-LISS, INC. CHAETODON LATEROPHYSIC CONNECTION 1339 TABLE 1. Acanthomorph families with representatives surfaces of the supracleithra’’ (Fig. 1). Webb and that have anterior swim bladder horns Blum (1990) and Webb (1998) further described this (modified from Smith, 2000) feature in Chaetodon using histological analysis, Reference calling it the ‘‘laterophysic connection’’ (LC) to draw attention to its apparent structural and putative Acropomatidae Katayama, 1959 Centropomidae Katayama, 1959 functional similarity to the simple otophysic connec- Chaetodontidae Gunther, 1860; Blum, 1988 tions in other fishes. Two LC types were described Cichlidae Dehadrai, 1959 (direct and indirect), based on whether or not the Ephippidae Herre and Montalban, 1927 swim bladder horns are in direct contact with the Gerreidae Green, 1971 Haemulidae Johnson, 1980 medial opening in the supracleithrum (Webb, 1998). Holocentridae Nelson, 1955 Among species with an indirect LC, most species Kuhliidae Gosline, 1966 have long horns, but a few species have short horns Lactariidae Leis, 1994 (Webb and Smith, 2000; Smith et al., 2003). Menidae Johnson, pers. commun. The two LC types found among Chaetodon species Moridae Parker, 1882 Moronidae Katayama, 1959 are analogous to the kind of variation found in the Nemastiidae Rosenblatt and Bell, 1976 simple otophysic connections among holocentrid Percichthyidae MacDonald, 1978 subfamilies (Nelson, 1955). Coombs and Popper Polyprionidae Katayama, 1959 (1979) demonstrated that Myripristis (subfamily Priacanthidae Starnes, 1988 Scombridae Collette and Nauen, 1983 Myripristinae), which has robust anterior swim Sciaenidae Sasaki, 1989; Chao, 1986, 1995 bladder horns that come into intimate contact with Sillaginidae McKay, 1985 the otic capsule (Nelson, 1955; unpub. data) and Sparidae Tavolga, 1974 modified inner ear morphology (Popper, 1977), dem- onstrates higher sensitivity to sound stimuli over a broader frequency range, when compared to Sargo- Popper, 1992). For instance, clupeiform fishes have centron (Adioryx, subfamily Holocentrinae), which complex otophysic connections in which anterior lacks swim bladder horns (Nelson, 1955; unpub. extensions of the swim bladder (otic bullae) invade observ.) and has unmodified inner ear morphology the otic capsule and directly contact thefluids of the (Popper, 1977). Ramcharitar et al. (2002, 2004) has inner ear and the cranial lateral line canals forming also demonstrated correlations among auditory the ‘‘recessus lateralis’’ (e.g., O’Connell, 1955; Best capabilities (thresholds and frequency response), and Gray, 1980). Otophysan fishes have modified and morphology of the swim bladder and ear among vertebral elements that form the Weberian appara- several genera of drums (family Sciaenidae). tus, which mechanically link the swim bladder with Webb (1998) suggested that the LC in Chaetodon the otic capsules (e.g., Alexander, 1962; Rosen and transmits pressure from the air-filled swim bladder Greenwood, 1970; Fink and Fink, 1996). to the fluid-filled lateral-line canal via the anterior Simple otophysic connections in which anterior swim bladder horns at the LC, initiating fluid move- swim bladder horns contact the otic capsule or invade ments in the lateral-line canal that are capable of the otic capsule (as bullae) are found in osteoglosso- stimulating canal neuromasts in the vicinity of the morphs and elopomorphs (Bridge, 1900; Greenwood, LC. She hypothesized that the presence of anterior 1970) and have been shown to enhance auditory capa- swim bladder extensions and a LC in Chaetodon bilities in mormyrids (Yan and Curtsinger, 2000; makes the lateral line sensitive to sound pressure. Fletcher and Crawford, 2001). Simple otophysic con- This would expand the functional repertoire of the nections have also been investigated experimentally lateral line system to include the reception of sound in a few acanthomorph taxa (holocentrids, Coombs pressure stimuli. The observation that swim bladder and Popper, 1979, and sciaenids, Ramcharitar et al., morphology is correlated with LC (Webb and 2002, 2004). Interestingly, anterior swim bladder Smith, 2000) suggests that swim bladder bioacoustics horns are known in representatives of at least 21 play an important role in LC function, thus demand- other acanthomorph families (Table 1), suggesting ing a closer examination of the swim bladder itself. that modifications of the swim bladder that enhance hearing may be even more widespread among fishes. In butterflyfishes of the genus Chaetodon,swimblad- Visualization of Swim Bladder Morphology der horns form an association, not with the otic cap- Reliable methods for assessing swim bladder mor- sule, but with a medial opening in the supracleithral phology are essential for an analysis of its bioacous- lateral line canal, defining a novel specialization, the tics (for both sound production and sound reception) laterophysic connection. and patterns of acoustic backscatter (for detection and identification of fish populations, Foote, 1980, The Laterophysic Connection 1985, 1988; MacLennon and Simmonds, 1992; Blum (1988: 121) diagnosed Chaetodon with Schaefer and Oliver, 2000; Foote and Francis, 2002). ‘‘bilaterally, paired, bulbous, antero-lateral divertic- However, descriptions of teleost swim bladders are ula, that are attached to the [incomplete] medial scattered throughout the ichthyology literature and

Journal of Morphology DOI 10.1002/jmor 1340 J.F. WEBB ET AL. assessments of the comparative morphology of the and Blaxter, 1989). This is not surprising because teleost swim bladder in taxonomic or phylogenetic swim bladder morphology is easily distorted by contexts are rare (e.g., Dobbin, 1941; Whitehead preparation and fixation artifact, such as the effects of rapid changes in hydrostatic pressure during the collection process, postmortem changes prior to fixa- tion, chemical fixation and storage in alcohol, and freezing and sectioning (Foote, 1985, 1988). Fur- thermore, when fishes are cleared and stained for osteological analysis (e.g., Pothoff, 1984), the swim bladder is frequently rendered transparent. Radio- graphic images of live, anesthetized or freshly fixed material, in which air is still retained in the pres- surized swim bladder, and into which liquid has not yet diffused, can provide only limited views of the swim bladder (e.g., Webb and Smith, 2000; Fig. 1). Computed X-ray tomography (CT) is a non-inva- sive imaging method that has revolutionized our ability to analyze the internal structure of living organisms. The high level of image resolution pro- vided by CT allows differentiation of the structure of tissues and materials of differing densities (e.g., soft tissue, bone, fluid and air). CT has proven to be particularly valuable for determining volumetric measures, imaging structurally complex organs, such as the inner ear of vertebrates (Ketten et al., 1998), and for the comparative osteological analysis of fossil and living vertebrates for systematic and functional studies (e.g., Witmer, 2001; Schaefer, 2003; Witmer et al., 2003; Summers et al., 2004; Kearney et al., 2005). The swim bladder, composed of soft connective tis- sues containing low density gases, is surrounded by dense muscle and bone, and is thus a prime candi- date for CT analysis (eg., see Carpenter et al., 2004). By visualizing the swim bladder in live, anesthe- tized fishes, CT imaging can allow the analysis of swimbladder structure without the introduction of the sorts of preparation artifact that arise when fixed specimens are examined.

Goals of the Current Study Webb and Smith (2000) presented preliminary data on LC morphology based on a histological anal-

Fig. 1. Camera lucida drawings of the dermal skeletal ele- ments just behind the orbit at the posterior margin of the skull in three species of Chaetodon with different LC morphologies. A: Chaetodon octofasciatus (AMNH 43117, LC Variant Dir1, see Table 3). Schematic representation of the anterior horn of the swim-bladder (shaded) that sits deep to the medial opening in the supracliethrum (black teardrop), the site of the LC. The supracleithral neuromast (gray oval) is just rostral to the LC, and the neuromasts in the first two lateral line scales (gray ovals) are just caudal to the LC (modified from Webb and Smith, 2000) B: Chaetodon multicinctus (AMNH 88343SW, LC Variant Ind2). C: Chaetodon ornatissimus (AMNH 88417SW, LC Variant Ind3). gb, swim bladder; h, horn; le, lateral extra- scapular; me, medial extrascapular; nm, neuromast; pt, post- temporal; pte, pterotic; s, supracleithrum. Scale bars ¼ 1 mm.

Journal of Morphology DOI 10.1002/jmor CHAETODON LATEROPHYSIC CONNECTION 1341 ysis of only eight species, but demonstrated that LC TABLE 2. Summary of material examined using histological morphology varies interspecifically and is not sexu- analysis, computed X-ray tomography (CT) and gross dissection ally dimorphic. Smith et al. (2003) identified several Species Histology CT Dissection morphological characters based on soft tissue mor- phology that were used to erect a new phylogenetic Non-Chaetodon Species rostratus 3 hypothesis of chaetodontid intrarelationships and nigrorostris 3 discussed the evolution of the laterophysic connec- Forcipiger flavissimus 33 tion. The purpose of this paper is to define interspe- Hemitaurichthys polylepis 3 cific variation in the morphology of the LC, swim Genus Chaetodon C. striatus 33 3 bladder and swim bladder horns among Chaetodon S.G. Chaetodon species in all Chaetodon subgenera using gross dis- C. capistratus 33 3 section, histological analysis and CT imaging. The C. humeralis 3 results of this work provide the morphological con- C. marleyi 3 33 text for the comparative functional analysis of the C. ocellatus S.G. Rabdophorus laterophysic connection. C. auriga 33 C. ephippium 33 C. falcula 3 MATERIALS AND METHODS C. fasciatus 3 C. flavirostris 3 Forty-one Chaetodon species (from all 11 Chaetodon subge- C. lineolatus 3 nera) as well as representatives of four other chaetodontid genera C. lunula 3 (Forcipiger, Hemitaurichthys, Heniochus, Johnrandallia) were C. melannotus 3 studied by analyzing cleared and stained specimens, histological C. rafflesi 3 material, X-ray and CT images, and dissected swim bladders (Ta- C. semilarvatus 3 ble 2). An IACUC-approved protocol was followed throughout. C. ulientensis 3 C. vagabundus 3 S.G. Roaops Osteological Analysis C. tinkeri 3 S.G. Exornator Several cleared and stained specimens (Chaetodon octofascia- C. citrinellus 3 tus, AMNH 43117; C. multicinctus AMNH 88343SW; C. ornatis- C. fremblii 3 simus AMNH 88417SW) (following methods in Pothoff, 1984) C. guttatissimus 3 and one dry skeleton (C. auriga MCZ 154429) were used to C. miliaris 33 describe the lateral line canals in the vicinity of the laterophy- C. multicinctus 33 sic connection, and were illustrated using a camera lucida or C. pelewensis 3 documented photographically. C. punctatofasciatus 3 C. quadrimaculatus 3 C. sendentarius 33 3 Histological Analysis of the S.G. Lepidochaetodon Laterophysic Connection C. kleinii 33 C. unimaculatus 33 Between one and eight individuals (juveniles and adults, 25– S.G. Megaprotodon 109 mm SL) in each of 22 Chaetodon species (from 10 of 11 subge- C. trifascialis 3 nera) as well as several individuals of Forcipiger flavissimus were C. oligacanthus 33 obtained commercially, through field collections in , or S.G. Gonochaetodon from museum collections, and prepared histologically (see Smith C. baronessa 33 et al., 2003). Voucher specimens of eleven species have been de- S.G. Tetrachaetodon posited in the Museum of Comparative (Harvard Univer- C. bennetti 3 sity, MCZ 156946-156959). Live fishes were anesthetized with an C. plebius 33 MS 222 solution in seawater until unresponsive. The body cavity S.G. Discochaetodon and orbits were injected with a solution of 10% formalin in sea- C. aureofasciatus 33 water and fishes were immersion fixed. All specimens were radio- C. octocasciatus 33 3 graphed within a day of fixation to visualize the air-filled swim C. rainfordi 3 bladder. Fish heads (of freshly fixed specimens, or in some cases, S.G. Corallochaetodon alcoholic museum specimens) were decalcified with Cal-Ex C. trifasciatus 33 3 (Fisher) overnight, or with 0.1 M di-NaEDTA in a solution of 10% S.G. Citharoedus formalin in seawater for 1–3 weeks with several solution C. meyeri 3 changes. Decalcification was confirmed radiographically. Fish C. ornatissimus 33 heads were trimmed, dehydrated in an ascending ethanol series, C. reticulatus 3 embedded in glycol methacrylate resin (Historesin [Leica], JB-4 [Polysciences], or Technovit 7100 [Kulzer, Wehrheim, Germany]) See text for details. and sectioned transversely at 5 lm with a tungsten-carbide knife on a Leica motorized microtome. Sections were mounted out of water onto chrom-alum subbed slides, dried overnight (608C), Measurements of lateral line canal diameter, horn diameter stained with 0.5% cresyl violet, air-dried overnight, and cover- and distance from the lateral line canal to the swim bladder horn slipped. Additional sections were mounted on positively charged were taken at the rostro-caudal midpoint of the medial opening of slides (Mt. Washington Scientific), stained with Sudan Black B the supracleithrum on the left and right sides in all specimens (Bennett et al., 1976) and coverslipped with glycerine in order to prepared histologically. In addition, the distance from the inner determine the distribution of adipose tissue in the vicinity of the wall of the swim bladder horn to the lumen of the otic capsule laterophysic connection. was taken in the caudal most section in which the lumen of the

Journal of Morphology DOI 10.1002/jmor 1342 J.F. WEBB ET AL. otic capsule was visualized. The rostro-caudal length and dorso- 30200 (n ¼ 1, 74 mm), ANSP 31576-91 (n ¼ 1, length unknown), ventral height of the medial opening in the supracleithrum and C. fremblii MCZ 2697 (n ¼ 1, 115 mm), C. guttatissimus ANSP the neuromasts in the vicinity of the LC (the post-temporal, 108362 (n ¼ 1, 73 mm), C. pelewensis, MCZ 82674 (n ¼ 1, 80 supracleithral canal neuromasts, and canal neuromasts in the mm), C. quadrimaculatus MCZ 16223 (n ¼ 1, 115 mm), ANSP first and second lateral line scales) were calculated in four species 97795 (n ¼ 1, 100 mm), C. sedentarius MCZ 59605 (105 mm). (Chaetodon octofasciatus, n ¼ 3, 63–73 mm SL; C. multicinctus, Subgenus Lepidochaetodon – Chaetodon unimaculatus MCZ n ¼ 3, 83–90 mm SL; C. kleinii, n ¼ 3, 74–85 mm SL; C. miliaris, 5744 (n ¼ 1, 115 mm). Subgenus Megaprotodon – Chaetodon oli- n ¼ 2, 45–50 mm SL; see Smith, 2000). Length was calculated gacanthus ANSP 100113 (n ¼ 1, length unknown). Subgenus by counting the number of sections in which tissue of interest Gonochaetodon – Chaetodon baronessa MCZ 64348 (n ¼ 1, 70 was present and multiplying by inter-section interval and mm). Subgenus Tetrachaetodon – Chaetodon bennetti MCZ 16260 section thickness. Width was measured in individual histological (n ¼ 1, 117 mm), C. plebius MCZ 64340 (n ¼ 1, 96 mm). Subgenus sections at the rostro-caudal midpoint of the neuromast or medial Discochaetodon – MCZ 64388 (n ¼ 1, supracleithral opening using SPOT1 software (Diagnostic Instru- 83 mm). Subgenus Corallochaetodon – Chaetodon trifasciatus ments, Sterling Heights, MI) and a drawing tablet. Measure- MCZ 89979 (n ¼ 1, 95 mm). Subgenus Citharoedus – Chaetodon ments taken in histological material are affected by shrinkage reticulatus MCZ 30210 (n ¼ 1, 108 mm). Other chaetodontid gen- and other preparation artifact and sectioning angle, so only sum- era – Chelmon rostratus, MCZ 46530 (n ¼ 1, 108 mm), Johnran- mary data are presented. dallia (Pseudochaetodon) nigrorostris, MCZ 40453 (n ¼ 1, 90 Specimens analyzed histologically. Fish length is ex- mm). pressed as standard length (SL). (Museum of Comparative Zool- ogy [MCZ], Harvard University, Cambridge; Academy of Natural Sciences, Philadelphia [ANSP]; Australian Museum, Sydney CT Imaging of Swim Bladder [AMS]). Webb Lab accession codes (e.g., Cc1) are retained for archival purposes. Genus Forcipiger – Forcipiger flavissimus (n ¼ 2, Computed X-ray tomographic (CT) images were obtained for 109 mm- SL [uncat. indiv., Ff1]). Genus Chaetodon: Incertae sedis – one to four individuals from each of 16 Chaetodon species (in 8 of Chaetodon striatus (n ¼ 1, 74 mm [Cs4]). Subgenus Chaetodon – 11 Chaetodon subgenera), as well as for three Forcipiger flavissi- Chaetodon capistratus (n ¼ 3, 25–52 mm [Cc1, 2, 4]), C. ocellatus mus and one Hemitaurichthys polylepis. Each fish was anesthe- (n ¼ 1, 41.8 mm [Cl7]). Subgenus Rabdophorus – Chaetodon tized in a seawater solution of MS 222 (Sigma) until ventilation auriga (n ¼ 1, 61.0 mm [Ca2]), C. ephippium (n ¼ 1, 51.0 mm ceased and the fish was incapable of maintaining an upright pos- [Ce1]), C. semilarvatus (n ¼ 1, 87 mm [Cv1]). Subgenus Exorna- ture. The fish was then placed left side up in a plastic box filled tor – Chaetodon miliaris (n ¼ 4, 45–100 mm [Cr1, 4, 9, 11]), C. with the anesthetic solution and wet gauze was placed either multicinctus (n ¼ 4, 83–90 mm, [Cm4, 5, 7, 9]). Chaetodon seden- beneath or on top in order to ensure flat and stable positioning tarius (n ¼ 1, 50 mm [Cd1]). Subgenus Lepidochaetodon – Chae- during CT scans. Following scanning, each specimen was over- todon kleinii (n ¼ 4, 70–85 mm [uncat. indiv., Ck10, 11, 13]), C. anesthetized in the anesthetic solution, injected with a solution unimaculatus (n ¼ 1, 38 mm [Cu1]). Subgenus Megaprotodon – of 10% formalin in seawater into the body cavity and orbits, and Chaetodon oligacanthus (n ¼ 1, 67 mm [Cg1 ¼ ANSP 100113]); immersion fixed for subsequent analysis. After initial scans, Chaetodon trifascialis (n ¼ 1, 87 mm, [Cz1]). Subgenus Gono- three individuals in two species (Chaetodon capistratus and C. chaetodon – Chaetodon baronessa (n ¼ 1, 55.0 mm [Cb1 ¼ AMS striatus) were immediately scanned a second time with the right 21915011]). Subgenus Tetrachaetodon – (n ¼ 2, side up, to identify any effects of scanning position on swim blad- 57–85 mm [Cp1 ¼ AMS 24678030, Cp2]). Subgenus Discochaeto- der morphology. don – Chaetodon aureofasciatus (n ¼ 1, 57.0 mm [Cq1 ¼ AMS CT scans were carried out on a Siemens Somatom Plus 4, a 24678030]), C. octofasciatus (n ¼ 8, 50–77 mm [Co9, 11, 12, 14, Siemens Volume Zoom (at the Radiology Department of the Mas- 20, 22, 30 þ 1 uncat. indiv.]), C. rainfordi (n ¼ 1, 71.7 mm [Ci1]). sachusetts Eye and Ear Infirmary, MEEI), or a Siemens Volume Subgenus Corallochaetodon – Chaetodon trifasciatus (n ¼ 1, 60 Zoom (at the Woods Hole Oceanographic Institution, WHOI). A mm [Ct1]). Subgenus Citharoedus – Chaetodon ornatissimus topogram (lateral view) was first obtained to visualize the full (n ¼ 2, 61–68 mm [Cn1, 4]), C. meyeri (n ¼ 1, 85.0 mm [Cy1]). extent of the swimbladder in each case. Scan regions were set to include the inner ear and the entire length of the swim bladder. Fish were scanned in the rostro-caudal axis using a spiral acqui- Gross Morphology of Swim Bladder and sition protocol at 200 mAs/140 kV (MEEI) or 150 mAs/120 kV Swim Bladder Horns (WHOI) using a 0.5 mm (WHOI) or 1.0 mm (MEEI) collimator width. Data were reformatted at 0.1 mm (WHOI) to 0.5 mm Conventional X-rays of all specimens were prepared prior to (WHOI and MEEI) to provide full high resolution series of trans- preparation of histological material (see above) and were exam- verse slices from which horizontal and sagittal MPR’s (multiple ined to determine overall swim bladder shape and position of the plane reconstructions) were reformatted with slice thicknesses of internal diaphragm formed by the tunica interna. In addition, 0.1 mm (WHOI) or 0.2 mm (MEEI). Three-dimensional recon- dissections were performed on one or more individuals in 30 structions were produced with a Shaded Surface Display (SSD) Chaetodon species (in 10 of 11 Chaetodon subgenera) and in one program (with Siemens proprietary software) using a range of X- species from each of three non-Chaetodon genera (Chelmon, ray attenuation values (1024/-510 Hu) appropriate for imaging Heniochus, Johnrandallia). the air within the swim bladder, the swim bladder horns and Museum material examined. Fish length is expressed as some details of the air–tissue boundary. standard length (SL). (MCZ, Museum of Comparative Zoology, Swim bladder and horn length (in the rostro-caudal axis, see Harvard University; ANSP, Academy of Natural Sciences, Phila- Table 4) were determined by counting the number of transverse delphia). Genus Chaetodon: Incertae sedis – Chaetodon striatus CT slices in which the swim bladder or swim bladder horns were MCZ 2710 (n ¼ 1, 112 mm). Subgenus Chaetodon – Chaetodon present and multiplying that value by slice thickness. Swim blad- capistratus MCZ 68477 (n ¼ 2, 82, 85 mm), C. humeralis MCZ der and horn diameters were measured with digital calipers on 43454 (n ¼ 1, length unknown), C. marleyi ANSP54806 (n ¼ 1, 90 large format CT films perpendicular to the dorso-ventral axis mm). Subgenus Rabdophorus – MCZ 16208 (defined by the median skeletal elements), at 1 or 2 mm intervals (n ¼ 2, 110, 112 mm), C. fasciatus MCZ 3705 (n ¼ 1, 125 mm), C. along the rostro-caudal length of the swim bladder, and at 0.5 flavirostris MCZ 36932 (n ¼ 1, 120 mm), C. lineolatus MCZ 82679 mm or 1 mm intervals along the length of the swim bladder (n ¼ 1, 80 mm), C. lunula MCZ 5751 (n ¼ 2, 110, 120 mm), C. mel- horns. annotus MCZ 6000 (C. dorsalis, n ¼ 1, 120 mm), C. rafflesi MCZ Material analyzed using CT. Fish length is expressed as 33168 (n ¼ 3, 70 mm), C. ulientensis MCZ 82677 (n ¼ 1, 125 mm), standard length (SL). MEEI, Massachusetts Eye and Ear Infir- ANSP86412 (n ¼ 1, 103 mm), C. vagabundus MCZ 30261 (n ¼ 3, mary; WHOI, Woods Hole Oceanographic Institution). Genus 82–87 mm). Subgenus Exornator – Chaetodon citrinellus MCZ Chaetodon: Incertae sedis – Chaetodon striatus (n ¼ 3, 61–70 mm

Journal of Morphology DOI 10.1002/jmor CHAETODON LATEROPHYSIC CONNECTION 1343 [Cs10, 11, 12], MEEI). Subgenus Chaetodon – C. capistratus (n ¼ 4; 3A,B). The supracleithral neuromast measures 65–77 mm [Cc 10, 11, 12, 13], MEEI), C. ocellatus (n ¼ 2, 52–73 300–700 lm (rostro-caudal axis) by 200–500 lm mm [Cl 40, 41], MEEI). Subgenus Rhabdophorus – Chaetodon auriga (n ¼ 2, 88–MEEI). Subgenus Roaops: C. tinkeri (n ¼ 2, (dorso-ventral axis). It is found on the medial wall 80–85 mm [Cx1, 2], WHOI). Subgenus Exornator – Chaetodon of the lateral line canal in the supracleithrum just miliaris (n ¼ 3, 70–90 mm [Cr 12, 15, 16], MEEI and WHOI), C. rostral to the medial opening in species with a direct multicinctus (n ¼ 4, 65–75 mm, [Cm10, 11, 12, 13], WHOI), C. LC (e.g., C. octofasciatus, Fig. 4B), or on the soft tis- ¼ ¼ punctofasciatus (n 1, 73 mm [Cf1], MEEI), C. sedentarius (n 1, sue that fills the medial opening in species with an 58 mm [Cd 7], MEEI). Subgenus Lepidochaetodon – C. kleinii (n ¼ 1, 70 mm [Ck16], MEEI). Subgenus Corallochaetodon – C. indirect LC (e.g., C. multicinctus, Fig. 4C). The first trifasciatus (n ¼ 1, 72 mm [Ct3], MEEI). Subgenus Citharoedus – two lateral-line scales just caudal to the medial C. ornatissimus (n ¼ 1, 65 mm [Cn 6], MEEI). Other genera: opening in the supracleithrum each contains a Forcipiger flavissimus (n ¼ 3, 90–114 mm, [Ff 7, 10, 11], MEEI ¼ somewhat smaller canal neuromast that is 200– and WHOI), Hemitaurichthys polylepis (n 1,110 mm [Hp1], 500 lm long (in its rostro-caudal axis) and 200– WHOI). l Two individuals from each of three species (Chaetodon capi- 400 m wide (in its dorso-ventral axis, Fig. 4A). stratus, C. ocellatus, C. striatus) were prepared histologically (see Methods above) several months after they were CT scanned and fixed in order to assess the degree of shrinkage and preparation Interspecific Variation in LC Morphology artifact present in histological material. Horn diameters in both CT images and histological material were measured in the medio- Variation in LC morphology is defined by the lateral (horizontal) axis at 180 lm intervals along their length proximity of swim bladder horns to the medial using an ocular micrometer mounted on a compound microscope. opening in the supracleithrum (direct/indirect LC), presence/absence of an external constriction (ductus communicans) in the tunica externa of the RESULTS swim bladder (one- or two-chambered swim blad- The configuration of the cranial lateral line canals der, sensu Dobbin, 1941), swim bladder horn width in the vicinity of the laterophysic connection (LC) is (wide/narrow) and length (long/short), and pres- demonstrated in cleared and stained material ence/absence of mucoid connective tissue in the soft (Fig. 1). A narrow, well-ossified lateral line canal tissue that fills the medial opening in the supra- (100–400 lm diameter) runs rostro-caudally cleithrum (characters 9, 36, 38, 39 and 40, respec- through the dorsal-most portion of the elongate tively, Smith et al., 2003). Two variants on a direct supracleithrum, the caudal-most element contain- LC (Dir1 and Dir2) and four variants on an indirect ing a portion of the cranial lateral line system. Cau- LC (Ind1-Ind4) were found among the 22 Chaeto- dally, the supracleithral canal is contiguous with don species analyzed histologically (Fig. 5). Only the trunk canal, which starts in the first lateral line one variant is present among species in each Chae- scale. Rostrally, the supracleithral canal connects to todon subgenus (as defined by Smith et al., 2003), a small canal segment in the post-temporal bone, with the exception of the Subgenus Citharoedus, in which is contiguous with the canal in the lateral which variants Ind3 and Ind4 are found (see Table extrascapular, and connects with the supratemporal 3). Four of the variants (Dir1, Ind1, Ind2 and Ind3) canal in the medial extrascapular and with the have been described briefly and illustrated sche- canal in the pterotic where the infraorbital, postotic, matically (Webb and Smith, 2000), but are more preopercular, and supraorbital canals meet. fully described and placed in a larger context below. The site of the LC is not a novel opening in the Variants on the direct laterophysic connec- medial wall of the supracleithrum (a ‘‘medial fossa,’’) tion. A direct LC is characterized by the direct con- as initially reported by Webb (1998), but is more tact of the swim bladder horn with the soft tissues appropriately interpreted as the posterior terminal filling the medial opening in the supracleithrum. pore of the lateral line canal segment in the supra- Two variants on the direct LC are defined here cleithrum (Smith et al., 2003). The supracleithrum (Table 3). is incomplete medially, forming an opening that Direct LC with mucoid connective tissue, wide ranges in shape from a rostro-caudally elongate oval horns, and a one-chambered swim bladder (Dir1) to a teardrop (Fig. 2). At the caudal end of the (Figs. 3A,B, 5A, 6A). A soft tissue ‘‘tympanum’’ supracleithral canal segment, the canal roof (the composed of four layers forms a 150–500 lm thick lateral wall of the canal) extends beyond the canal barrier between the fluid-filled lateral line canal floor (the medial wall of the canal), so that the pos- and the gas-filled swim bladder horn (Fig. 3B). The terior terminal pore of the canal segment in the thin epithelial lining of the lateral line canal is in supracleithrum points medially (arrows in Fig. 2). direct contact with the mucoid connective tissue, In Chaetodon octofasciatus, C. multicinctus and which stains pink with cresyl violet, and may con- C. kleinii the medial opening in the supracleithrum tain fat cells (as indicated by positive Sudan Black measures 500–800 lm in the rostro-caudal axis B staining (data not shown), and sits deep to the and 300–700 lm in the dorso-ventral axis. In medial opening the supracleithrum. Variation in smaller individuals of C. miliaris, the medial open- the number and distribution of fat cells may reflect ing measures only 300 lm in the rostro-caudal differences in the nutritional state among specimens axis and 500 lm in the dorso-ventral axis (Fig. examined. The collagenous tunica externa of the

Journal of Morphology DOI 10.1002/jmor 1344 J.F. WEBB ET AL. swim bladder horn is thinned in the vicinity of the medial opening and tightly adheres to the mucoid connective tissue (Figs. 3B, 6A,B). The tunica interna appears to be composed of multiple layers of very thin epithelium and readily separates from the tunica externa during histological preparation. In Chaetodon octofasciatus and C. rainfordi, the swim bladder horns overlap the inner ear (the lage- nar and saccular otolithic organs) in the rostro-cau- dal axis by <1 mm; overlap is not observed in other species with this LC variant. The swim bladder horns sit lateral to the neurocranium, and kidney tissue is generally found between the otic capsule and the horns at the level of the medial opening in the supracleithrum. The otic capsule is well-ossi- fied and does not appear to demonstrate any thin- ning or other structural modifications (Fig. 3A). The rostral end of the swim bladder horns extends to the level of the post-temporal (just rostral to the Fig. 2. Medial opening in the supracleithrum (arrows), site of the supracleithrum) where the saccular otolithic laterophysic connection (LC). A: Medial surface of supracleithrum organs and the vertical semicircular canals are also in dry skeleton of Chaetodon auriga (MCZ 154429). B: Lateral (left) visible in transverse sections. and medial (right) surfaces of the supracleithrum in Chaetodon sp. Direct LC without mucoid connective tissue and with narrow horns and a two-chambered swim supracleithrum. The epithelial lining of the lateral- bladder (Dir2) (Figs. 5B, 6B). Mucoid connective line canal, the tunica externa, and tunica interna tissue is not present in the medial opening of the form a thin tympanum between the narrow gas-

Fig. 3. Histological sections and CT images of the laterophy- sic connection (LC) and swim bladder. A: Transverse section through LC in Chaetodon octofas- ciatus.Scalebar¼ 500 lm(modi- fied from Webb, 1998). B: Close- up of laterophysic tympanum in another specimen of C. octof- asciatus. Scale bar ¼ 200 lm. C: Three-dimensional reconstruc- tion (CT) of air volume in the swim bladder and swim bladder horns in C. ephippium.Trailof small air bubbles are trapped in soft tissue (unknown origin), but are not associated with the swim bladder. D: Transverse CT slice at level of arrow 1 in B. E: Trans- verse CT slice at level of arrow 1 in C. Scale bar in D and E ¼ 10 mm. cns, central nervous system; h, horn; ie, inner ear; ll, lateral line canal; mct, mucoid connective tissue; s, supracleithrum; te, tun- ica externa; ti, tunica interna. (Re- produced with permission from Smith et al., Cladistics, 2003, 19, 287-306, ÓBlackwell Publishing).

Journal of Morphology DOI 10.1002/jmor CHAETODON LATEROPHYSIC CONNECTION 1345

Fig. 4. Canal neuromasts in the vicinity of the laterophy- sic connection (LC) in Chaetodon spp. A: Canal neuromast in lateral line scale (arrow heads) in C. multicinctus (Cm4). B: Neuromast caudal to the direct LC in C. octofasciatus (uncat. spec.). C: Neuromast in the supracleithrum at the level of the indirect LC (arrows) in C. multicinctus (Cm4). ft, fatty tissue; m, epaxial musculature; sm, stratum compactum; n, neuromast ; s, supracleithrum. Scale bars ¼ 200 lm.

Fig. 5. Schematic representation of the six variants on laterophysic connection (LC) morphology (in dorsal view) among Chaeto- don species (See Table 2). A: Dir1 – direct LC with mucoid connective tissue, wide horns, and one-chambered swim bladder (e.g., Chaetodon octofasciatus). B: Dir2 – direct LC without mucoid connective tissue, with narrow horns, and a two-chambered swim bladder (e.g., C. plebeius). C: Ind1 – indirect LC with mucoid connective tissue, wide horns and one-chambered swim bladder (e.g., C. kleinii). D: Ind2 – indirect LC without mucoid connective tissue and with wide horns and one-chambered swim bladder (e.g., C. multicinctus). E: Ind3 – indirect LC with short horns and a one-chambered swim bladder (e.g., C. ornatissimus). F: Ind4 – indirect LC with short horns, a one-chambered swim bladder and a medial anterior extension of swim bladder (e.g., C. meyeri). d, trans- verse diaphragm in tunica interna; h, swim bladder horn; hc, sensory macula; ie, inner ear in otic capsule; kt, kidney tissue; ll, 1st lateral-line scale; m, muscle; mct, mucoid connective tissue; nm, neuromast; s, supracleithrum; te, tunica externa (solid line); ti, tunica interna (dotted line).

Journal of Morphology DOI 10.1002/jmor 1346 J.F. WEBB ET AL.

TABLE 3. Morphological features of Chaetodon spp. with different LC variants as determined by histological analysis (MCT), gross dissection (SB type, No. chambers) and CT imaging (horn length and diameter)

MCT present Horn length Horn diameter No. of SB chambers SB type

Variant Dir1 (Incertae sedis and Subgenera Rabdophorus, Gonochaetodon, Discochaetodon) Chaetodon striatus yes long wide 1 free Chaetodon auriga yes long wide 1 free Chaetodon ephippium yes long wide 1 free Chaetodon baronessa yes long wide 1 free Chaetodon aureofasciatus yes long wide 1 free Chaetodon octofasciatus yes long wide 1 free yes long wide 1 ? Chaetodon semilarvatus ?a long ? 1 free Variant Dir2 (Subgenera Megaprotodon, Tetrachaetodon) Chaetodon trifascialis no long narrow 2 free Chaetodon plebeius no long narrow 2 free Chaetodon oligacanthus ? long narrow 2 free Variant Ind1 (Subgenera Lepidochaetodon) Chaetodon kleinii yes long wide 1 attached Chaetodon unimaculatus yes long wide 1 attached Variant Ind2 (Subgenera Chaetodon, Exornator, Corallochaetodon) Chaetodon capistratus no long wide 1 free Chaetodon ocellatus no long wide 1 free Chaetodon miliaris no long wide 1 attached Chaetodon multicinctus no long wide 1 attached Chaetodon sedentariusb no long wide 1 attached Chaetodon trifasciatus no long wide 1 attached Variant Ind3 (Subgenera Citharoedus, Roaops) Chaetodon ornatissimus no short N/A 1 attached Chaetodon tinkeri ? short N/A 1 ? Variant Ind4 (Subgenus Citharoedus) Chaetodon meyeri no short N/A 1 attached

MCT, mucoid connective tissue in medial opening of supracleithrum; No. SB Chambers, number of swim bladder chambers formed by tunica externa; SB type, morphology of swim bladder with respect to relationship to peritoneum. aCould not be determined because of the poor quality of histological tissue. bNot variant Ind1, as reported by Webb (1998). See text and Figures 5 and 6 for additional explanation.

filled swim bladder horn and the fluid-filled lat- 1 mm). Four variants on the Indirect LC are eral-line canal. In some specimens, adipose cells described here (Table 3). sit deep to the epithelial lining of the lateral line Indirect LC with mucoid connective tissue, long, canal.AttheleveloftheLC,kidneytissuesits wide horns, and a one-chambered swim bladder between the otic capsule and the swim bladder (Ind1) (Figs. 5C, 6C). Epaxial muscle tissue sits horns, which appear to be firmly attached to the between the supracleithral lateral line canal and medial surface of the supracleithrum. In one spec- the swim bladder horn, precluding direct contact imen of Chaetodon plebeius, the swim bladder between them. Adipose cells may be found within horns appear to bulge out into the supracleithrum the mucoid connective tissue that fills the medial through the medial opening. This may be the opening in the supracleithrum. The wide horns result of preparation artifact, but this was not originate in a common trunk at the rostral end of confirmed with CT. In both C. trifascialis and C. the body of the swim bladder and appear uniform plebeius, the swim bladder horns extend rostrally, in diameter with a tapered rostral end. to the level of the otic capsule. The prominent con- Indirect LC without mucoid connective tissue, striction in the tunica externa is located about with long, wide horns and a one-chambered swim 20–25% down the length of the swim bladder, bladder (Ind2) (Figs. 5D, 6D). Mucoid connective forming a two-chambered swim bladder (described tissue is not found in the LC, but most of the speci- by Blum, 1988) in C. oligacanthus and C. trifas- mens with this variant have adipose cells between cialis). the epithelial lining of the lateral line canal and Variants on the indirect laterophysic con- the epaxial musculature deep to the LC. It is inter- nection. An indirect LC is defined by the absence esting to note that Bauchot et al. (1989) did not of direct contact between the swim bladder horns mention the presence of the swim bladder horns in and the medial opening in the supracleithrum. their histological analysis of Chaetodon trifascia- Muscle, kidney and other soft tissues lie deep to tus, but this may be explained by the fact that their the medial opening, such that the distance between data were taken from a 16-mm-individual (prob- the lumen of the swim bladder horn and the lateral ably at tholichthys stage) in which the swim blad- line canal in the supracleithrum is variable (0.2– der horns had not yet developed.

Journal of Morphology DOI 10.1002/jmor CHAETODON LATEROPHYSIC CONNECTION 1347

Fig. 6. Variants on the laterophysic connection (LC) among Chaetodon species represented schematically in Figure 5. A: Dir1 – direct LC with mucoid connective tissue in Chaetodon octofasciatus (uncat. spec.). Scale bar ¼ 500 lm. B: Dir2 – direct LC without mucoid connective tissue in Chaetodon plebeius (Cp1) Scale bar ¼ 200 lm. C: Ind2 – indirect LC with no mucoid connective tissue in Chaetodon multicinctus (Cm4). Scale bar ¼ 500 lm. D: Ind1 – indirect LC (arrows) with mucoid connective tissue in Chaetodon kleinii (Ck10). Scale bar ¼ 200 lm. E: Ind3 – indirect LC with short horns in Chaetodon ornatissimus (Cn4). Scale bar ¼ 500 lm. h, swim bladder horn; kt, kidney tissue; ll, lateral-line canal; mct, mucoid connective tissue; m, muscle; n, neuromast; s, supra- cleithrum; te, tunica externa; ti, tunica interna (Reproduced with permission from Smith et al., Cladistics, 2003, 19, 287–306, Arrows in A, C, and E indicate the medial opening in the supracleithrum. Ó Blackwell Publishing).

Indirect LC without mucoid connective tissue, the medial opening in the supracleithrum as in spe- with short horns and a one-chambered swim blad- cies with long horns (Figs. 5E, 9G,H). Mucoid connec- der (Ind3) (Figs. 5E, 6E). The short swim bladder tive tissue is absent at the level of the supracleith- horns extend <1mmfromtherostralendofthebody rum. Epaxial musculature is present deep to the of the swim bladder and do not extend to the level of supracleithrum as in other species with an indirect

Journal of Morphology DOI 10.1002/jmor 1348 J.F. WEBB ET AL.

Fig. 7. CT images demonstrating rela- tionship of the swim bladder, swim bladder horns and skeleton in Chaetodon punctofas- ciatus. A: Parasagittal CT slice demonstrat- ing vertical orientation of swim bladder horn (h) lateral to vertebral column. B: Midsagittal CT slice demonstrating the close spatial relationship of the dorsal mid- line of the swim bladder and vertebral col- umn (white). C: Horizontal CT slice demon- strating the position of the ribs (arrows) against the fully inflated swim bladder. D: 3D reconstruction of the volume of air within the swim bladder and swim bladder horns. Periodic indentations in lateral sur- face of gas volume correspond to the posi- tion of ribs illustrated in C.

LC and kidney tissue is present deep to the supra- as a depression in 3D reconstructions of the volume cleithrum where the long swim bladder horns are of air within the swim bladder (Figs. 3, 9). An found in other Chaetodon species (Fig. 7E). infolding of the tunica interna divides the swim Indirect LC without mucoid connective tissue, bladder lumen transversely, forming a translucent with short horns, medial rostral extension and a diaphragm, with a small (1–3 mm) central open- one-chambered swim bladder (Ind4) (Fig. 5F). This ing that subdivides the lumen of the swim bladder variant is also characterized by short swim bladder into anterior and posterior compartments (the horns (see Variant Ind3, Chaetodon ornatissimus), ‘‘chambers’’ of some authors; Fig. 8B,C Smith et al., but there is an additional median anterior exten- 2003). The location of the diaphragm is variable sion of the swim bladder, which extends rostrally to among individuals examined (50–75% down the the level of the supracleithrum. length of the swim bladder) and it can be visualized in X-rays (Fig. 8A), as well as in some CT slices. In those species with two-chambered swim bladders (Variant Dir2, Chaetodon oligacanthus, C. trifascia- Morphology of the Swim Bladder lis, and C. plebeius), this diaphragm is located cau- and Swim Bladder Horns dal to the constriction in the tunica externa (the Gross dissection, histology, X-rays and CT imag- ductus communicans). ing (2D transverse, sagittal and horizontal slices Transverse CT slices demonstrate that the swim and 3D reconstructions) reveal details of the mor- bladder horns in Chaetodon are roughly circular in phology of the physoclistous swim bladder and crossection, but in species with a direct LC, the lat- swim bladder horns (Figs. 3, 7, 8). In all species eral surface of the horn appears somewhat flat- examined, the swim bladder lies in the dorso- tened at the point of contact with the medial open- medial region of the body cavity just ventral to the ing of the supracleithrum (Fig. 3). Three-dimen- vertebral column and lacks any intrinsic or extrin- sional reconstructions (CT) demonstrate that in sic musculature that might be indicative of a swim some species the horns have a slightly narrower bladder mechanism of sound production (Fig. 7). A neck with an expanded distal end (as per Blum, thick ventral midrib runs along the length of the 1988) and small depressions in the surface of the tunica externa of the swim bladder and is visible horns that are not evident in the quantitative anal- upon gross dissection, in histological sections, and ysis of two-dimensional CT images (Figs. 3, 7, 9).

Journal of Morphology DOI 10.1002/jmor CHAETODON LATEROPHYSIC CONNECTION 1349 2000). Species with a direct LC (e.g. C. octofascia- tus) have a swim bladder with a distinct change in angle in its long axis (tilt angle), such that the an- terior half of the swim bladder has a relatively ros- tro-caudal orientation relative to the posterior half of the swim bladder, which has a more dorso-ven- tral orientation (Figs. 3C, 8A). The swim bladder is robust with a thick tunica externa composed of what appear to be multiple layers of collagen. A thin, translucent, lightly pigmented peritoneum covers the ventral surface of the swim bladder, and wraps tightly around its posterior end, so that the swim bladder is clearly visible upon lateral dissec- tion of the abdominal cavity. The bilateral swim bladder horns extend separately from the dorso- rostral surface of the swim bladder. This swim bladder morphology is defined as a ‘‘free’’ swim bladder to reflect the fact that the posterior end of the swim bladder appears to sit free in the perito- neal cavity. In Chaetodon species with an indirect LC, the swim bladder appears to be more smoothly con- toured (lacking an abrupt change in angle), and may decrease in circumference caudally (Figs. 7, 9). The relatively translucent tunica externa appears Fig. 8. Internal transverse diaphragm in the swim bladder of Chaetodon spp. A: X-ray showing the swim bladder (white) and to be thinner dorsally than it is ventrally, and the transverse diaphragm (arrow) in C. octofasciatus. B: Central much thinner overall when compared to the swim opening in isolated diaphragm of C. octofasciatus (2–3 mm diam- bladder in species with a direct LC. The swim blad- eter). C: Diaphragm in ventral view in intact swim bladder of der sits above a thick, opaque peritoneum, which C. multicinctus with anterior end removed (rostral is to left). attaches laterally to the ribs so that the swim blad- der is not visible upon lateral dissection of the ab- Two swim bladder morphologies are found among dominal cavity. The swim bladder horns extend ros- Chaetodon spp. and are generally correlated with trally in a dorso-ventrally flattened common trunk LC type (Table 3; Webb, 1998; Webb and Smith, that divides into the two horns. This is defined as

Fig. 9. Three-dimensional reconstructions and parasagittal CT images of swim bladder. A/B: Forci- piger flavissimus. Note absence of horns and lateral compression of gas volume. C/D: Chaetodon punc- tofasciatus. E/F: Chaetodon miliaris. G/H: Chaetodon ornatissimus. Note presence of short horns. I/J: Chaetodon kleinii, K/L: Chaetodon trifasciatus. Artifactual outpocketing of swim bladder can be seen on ventral surface of swim bladder.

Journal of Morphology DOI 10.1002/jmor 1350 J.F. WEBB ET AL.

TABLE 4. Quantitative analysis of swim bladder and horn morphology as revealed by CT imaging

Max. horn Horn length diameter LC Length SB SB/SL Max. SB Species variant (SL) length length diameter RL R L

C. auriga Dir1 88 27.5 0.31 7.81 3.5 3.5 2.21 2.40 115 34.5 0.30 10.04 5.5 5.5 3.44 3.32 C. striatus Dir1 60Aa 15.0 0.25 8.52 2.5 2.5 2.37 2.36 60Ba 15.0 0.25 6.11 3.0 2.5 3.18 2.06 61a 14.0 0.23 6.36 2.0 2.0 1.99 2.15 61Ba 12.0 0.20 6.13 2.0 2.0 2.02 2.53 70 15.0 0.21 6.41 2.5 2.5 2.19 2.14 C. capistratus Ind2 65 20.5 0.32 7.69 3.5 2.5 2.42 2.51 66 20.0 0.30 7.76 2.5 3.0 2.41 2.59 75 22.5 0.30 8.51 3.5 3.0 2.79 8.15 77Aa 24.5 0.32 9.11 3.5 4.0 2.81 3.36 77Ba 22.0 0.29 8.73 3.5 3.5 2.65 2.72 C. miliaris Ind2 85 26.5 0.31 7.87 3.0 3.0 2.95 2.90 90 29.0 0.32 8.55 3.5 3.5 3.26 3.21 C. multicinctus Ind2 58 18.0 0.31 5.88 2.5 2.0 2.31 2.36 65 19.0 0.29 6.54 2.0 2.0 2.28 2.36 68 21.5 0.32 6.24 3.0 3.0 2.38 2.57 75 22.5 0.31 7.12 3.5 3.5 2.95 2.96 C. ocellatus Ind2 52 13.0 0.25 5.44 1.5 1.5 1.07 1.27 73 21.0 0.29 6.95 3.0 3.0 2.69 1.86 C. sedentarius Ind2 58 18.5 0.32 6.35 2.5 2.5 1.99 2.17 Forcipiger – 103 27.5 0.27 8.20 – – – – flavissimus – 114 29.5 0.26 8.46 – – – – Hemitaurichthys – 110 30.5 0.28 10.57 – – – – polylepis aA, fish scanned first with left side up; B, fish scanned subsequently, with right side up. All measurements in mm, with measurement error ¼ 60.5 mm for swim bladder (SB) and horn length. See text for methodological details and Table 3 for definition of LC variants. an ‘‘attached’’ swim bladder to reflect the fact that were easily derived from transverse CT slices at least part of the swim bladder is attached to the (Table 4). The swim bladder in Chaetodon spp. is peritoneal lining, and that the swim bladder does 12–34 mm long (20–32% of SL) with a maximum not appear to sit free in the abdominal cavity. diameter of 10 mm. The swim bladder horns are Gross dissection of 30 Chaetodon spp. (in 10 of 12 1.5–3.5 mm long, and generally 2–3 mm in diame- Chaetodon subgenera, Table 1) revealed that, with ter. Swim bladder length, as well as maximum the exception of one species in subgenus Exornator swim bladder and swim bladder horn diameter, is and two species in subgenus Chaetodon (C. capi- generally correlated with fish size in those Chaeto- stratus and C. ocellatus; see Table 3), each species don species for which more than one individual has the same swim bladder morphology as others was examined. Measurements of horn length and in the same subgenus in which LC morphology had diameter do not suggest any consistent right-left already been determined histologically. Thus, gross asymmetry (Table 4). morphology of the swim bladder (free/attached) While it appears that horn length did not change in appears to be a reliable predictor of LC morphology either Chaetodon capistratus or C. striatus in se- (direct/indirect) among Chaetodon species. quential CT scans (see Table 4), swim bladder length Swim bladder morphology in non-Chaetodon gen- and maximum swim bladder diameter did decrease in era appears to be similar to that in Chaetodon spe- specimens of both species. Interestingly, while maxi- cies with an indirect LC. In Forcipiger, the swim mum horn diameter appeared to decrease in C. capi- bladder is elongate and sausage-shaped, but is stratus (indirect LC), it appeared to increase in C. more laterally compressed, giving the swim bladder striatus (a direct LC) in a subsequent scan. an oval cross-section along most of its length (Fig. Some shrinkage and preparation artifact was evi- 9A,B). Gross dissection demonstrates that Henio- dent in all histological material examined subse- chus has a swim bladder similar in to those Chaeto- quent to CT scanning (Table 5). Horn diameter was don species with an indirect LC. CT scans of Hemi- compared in CT images and histological sections taurichthys reveal that it has a swim bladder with prepared from the same specimens. Chaetodon a gradual contour, suggesting that its swim bladder striatus (direct LC), clearly demonstrates shrink- is similar to that in the other non-Chaetodon spe- age (histology/CT ratio ¼ 0.5–0.6) while C. capistra- cies examined. tus and C. ocellatus (indirect LC) did not (histology/ Swim bladder and swim bladder horn dimensions CT ratio ¼1.0). One specimen of C. ocellatus had in nine Chaetodon and two non-Chaetodon species a high ratio (1.6) and a horn diameter that is less

Journal of Morphology DOI 10.1002/jmor CHAETODON LATEROPHYSIC CONNECTION 1351

TABLE 5. Comparison of measurements from CT images and histological material of right horn diameter (in mm) in three Chaetodon spp. Ratio of measurements from histological material and CT images (Histo/CT) are to the nearest 0.5. See Table 3 for definition of LC variants

Mean horn Mean horn Mean horn Max horn Max horn Horn LC Length diameter diameter diameter diameter diameter diameter Species variant (SL) (CT) (histo) (histo/CT) (CT) (histo) (histo/CT)

C. striatus Dir1 60 1.78 0.81 0.5 2.37 1.2 0.5 61 1.77 0.96 0.5 1.99 1.26 0.6 C. capistratus Ind2 66 1.85 1.50 0.8 2.41 2.06 0.9 77 2.41 2.82 1.2 2.81 2.69 1.0 C. ocellatus Ind2 52 1.04 1.06 1.0 1.07 1.71 1.6 73 1.65 2.03 1.2 2.69 2.56 1.0

than half that of the other specimens, which was known for their extensive and diverse social behav- likely due to the deflation of the horn. iors that involve close interactions among individu- als (Reese, 1975; Hourigan, 1989). The recent dis- covery of sound production by Forcipiger and Chae- Relationship of Swim Bladder and Horns todon species in the context of social behavior to the Axial and Cranial Skeleton (Tricas et al., 2004; Tricas and Boyle, 2005) demon- CT images demonstrate an intimate relationship strates the importance of acoustic stimuli to the between the swim bladder to the ribs and the verte- behavior and ecology of these fishes. The high noise bral column (Figs. 7, 9), features that cannot be levels on coral reefs make the extraction of biologi- appreciated in conventional X-rays. Rib indenta- cally significant acoustic stimuli particularly chal- tions observed on the outer surface of the tunica lenging. It will be important to measure ambient externa in fixed specimens examined grossly are noise levels within the frequency spectrum of but- not the result of fixation artifact. The close rela- terflyfish hearing capabilities to determine if this is tionship of the ribs to the fully inflated swim blad- indeed the case. However, chaetodontid species that der is shown in horizontal sections and in 3D recon- exhibit monogamous pairing behavior spend a great structions of the air volume that fills the swim deal of time in close proximity to one another (Tricas bladder, in which indentations of the ribs are visi- and Boyle, 2005), so that both the ear and lateral ble on the lateral surface of the air volume filling line system, which both respond to ‘‘near field’’ the swim bladder (Figs. 7C,D, 9). The swim bladder acoustic stimuli may be used to analyze signals pro- sits up against the vertebral column along most of duced during behavioral interactions. its length; posteriorly the dorsal surface of the Recent experiments have demonstrated that fill- swim bladder abuts the postabdominal vertebrae, ing of the swim bladder horns or disruption of the forming a distinct groove in the posterior-dorsal LC changes responses to sound stimuli and alters surface of the swim bladder (Fig. 9). social behavior (Tricas and Boyle, 2005). These The body of the swim bladder extends rostrally observations all lend support to the hypothesis that to a point just caudal to the well-ossified otic capsu- the swim bladder horns and LC are an adaptation les. The swim bladder horns extend rostrally by for reception of sound pressure by the lateral line another 2–3 mm and are positioned dorso-lateral system and/or the ear (Webb, 1998; Webb and to the otic capsule (Table 3). Variation in the angle Smith, 2000; Smith et al., 2003). Thus, we hypothe- of the major axis of the horns among species (from size that the simultaneous input of pressure stimuli primarily rostral to primarily dorsal) is easily from the swim bladder to both the lateral line and visualized in parasagittal CT slices and in lateral inner ear, as well as direct stimulation of the ear views of 3D reconstructions (Figs. 7, 9). In trans- (whole body acceleration), and hydrodynamic stimu- verse CT slices, which illustrate the medio-lateral lation of neuromasts in the lateral line canals of the plane, the swim bladder horns come within 500– head and trunk facilitates a unique interaction 1500 lm of the caudal end of the well-ossified otic among different acousticolateralis inputs (Braun capsule. et al., 2002) that may enhance the ability of Chaeto- don species to intepret behaviorally important acoustic stimuli in noisy environments. We further suggest that variation in LC mor- DISCUSSION phology among Chaetodon species may represent The laterophysic connection (LC), the association alternative adaptive strategies for the transmis- of anterior swim bladder horns with the medial sion of acoustic signals generated during social openings in the supracleithra, evolved in a single interactions in species that live in noisy coral reef clade of chaetodontid fishes. Butterflyfishes are well environments. If swim bladder horns in Chaetodon

Journal of Morphology DOI 10.1002/jmor 1352 J.F. WEBB ET AL. transduce a sound pressure wave-induced oscilla- swim bladder horn, will attenuate quickly as it tion of the gas volume in the swim bladder into moves through adjacent tissues. This may movements of the tunica externa of the swim blad- explain why the few Chaetodon species that have der and horn walls that generate a secondary par- narrow horns have a direct LC with no mucoid ticle displacement field, then we predict that connective tissue (Dir2, subgenera Megaproto- don, Tetrachaetodon, Fig. 6B), where the dis- 1. Swim bladder horn length will influence the tance between the wall of the swim bladder horn sensitivity of the ears to sound pressure. Fol- and the lumen of the lateral line canal is mini- lowing results from holocentrids (Coombs and mized. Popper, 1979) and sciaenids (Ramcharitar et al., 4. The presence of mucoid connective tissue in the 2002) the long swim bladder horns in the vicin- medial opening of the supracleithrum increases ity of the otic capsule, found in most Chaetodon the distance from the air in the horn lumen to species, should increase the amplitude of sound the fluid in the lateral line canal lumen, and it pressure-induced stimuli that reaches the ears. is predicted that the addition of this specialized Thus, Chaetodon species with long horns should tissue to the laterophysic tympanum will also have lower hearing thresholds when compared affect its biomechanical properties and thus the to Chaetodon species with short horns (regard- way in which the tympanum responds to acous- less of the proximity of the swim bladder horns tic stimuli. The fat content of laterophysic tym- to the medial openings in the supracleithra, LC panum was variable among individuals and is type) and to non-Chaetodon species that lack likely influenced by their nutritional state, horns altogether. which has a seasonal component (Figs. 11 and 2. The proximity of the long swim bladder horns 12 in Tricas, 1986). The functional significance to the medial opening in the supracletihrum of this fat is unknown, but the acoustic proper- (direct versus indirect LC) will influence the ties of fat have been investigated with respect ability of sound pressure stimuli to be transmit- to sound transmission to the ear of cetacea ted to the lateral line system. Therefore, in spe- (Wartzok and Ketten, 1999). cies with a direct LC (variants Dir1, Dir2), sound pressure stimuli will be transmitted more effectively to the lateral line canal system than Finally, the correlation of swim bladder morphol- in species with an indirect LC (variants Ind1, ogy (free/attached) and LC type (direct/indirect) Ind2) in which stimuli must move through mus- among Chaetodon species highlights the functional cle or kidney tissue in order to reach the medial importance of the swim bladder as a component of opening in the supracleithrum (Webb and the LC system and demands investigations of the Smith, 2000). In all Chaetodon species, in order biomechanics and bioacoustics of the swim bladder. for stimuli to be transmitted from swim bladder CT imaging provides direct evidence for the hetero- horn to the ear it must traverse soft tissue and geneity of the tissues immediately surrounding the the well-ossified bone of the otic capsule, swim bladder (peritoneum, ribs, musculature, also regardless of LC type. see images of fishes on www.digimorph.org), which 3. Horn diameter (wide/narrow) will determine the need to be considered when modeling the acoustic characteristics of the sound stimulus delivered to responses of the swim bladder (discussed by Sand both the lateral line system and ears. It has been and Hawkins, 1973; Foote and Francis, 2002). Pre- argued (see Schellart and Popper, 1992) that the dictions of swim bladder acoustic responses have displacement of the apex of the swim bladder in generally been made using linear and volumetric response to a sound stimulus is higher than at measures and models that describe the swim blad- the lateral swim bladder walls. Schellart and der as a sphere or prolate spheroid. More complex Popper (1992) predicted that the magnitude of models are needed to take into consideration the displacement of the swim bladder apex and the anatomical features of swim bladders. CT imaging resultant transmission of sound in surrounding of live, anesthetized fishes (even at resolutions tissues is related to swim bladder axis ratio (di- lower than one can achieve using high energy, high ameter/rostro-caudal length). Following their resolution CT, www.digimorph.org) offers an oppor- line of argumentation, it is predicted that the dis- tunity to accurately describe and quantify swim placement amplitude of the relatively sharp apex bladder morphology among a diverse range of spe- of swim bladder horns in response to a sound cies, thus facilitating comparative studies of its bio- pressure wave will be higher than that of the acoustic properties. blunt anterior wall of the body of the swim blad- der, and that the displacement amplitude of the wall of narrow horns will be greater than that in ACKNOWLEDGMENTS wide horns. They also indicate that the high am- plitude stimulus emerging from the sharp apex The authors thank Drs. Melanie Stiassny and of an elongated swim bladder, or in this case a Barbara Brown (AMNH) who provided a loan of

Journal of Morphology DOI 10.1002/jmor CHAETODON LATEROPHYSIC CONNECTION 1353 cleared and stained material. Dr. Dominique Daget Braun CB, Coombs S, Fay RR. 2002. What is the nature of mul- and Bill Saul (ANSP), and Dr. Jeff Leis and Mark tisensory interaction between octavolateralis sub-systems? Brain Behav Evol 59:162–176. McGrouther (AMS) provided loans or gifts of pre- Bridge TW. 1900. The air-bladder and its connection with the served materials. Many thanks to Karsten Hartel auditory organ in Notopterus borneensis. J Linn Soc Lond (MCZ, Harvard U.) and Dr. John Lundberg 27:503–540. (ANSP) for special permission to dissect material Carpenter KE, Berry TM, Humphries JM. 2004. Swim bladder in their collections. Brian DuVall (New Jersey and posterior lateral line nerve of the nursery fish Kurtus gulliveri (Perciformes: Kurtidae). J Morphol 260:193–200. State Aquarium) and Carl Meyer (U. Hawaii) Chao NL. 1986. A synopsis on zoogeography of the Sciaenidae. helped us to obtain specimens for this study. They In: Uyeno T, Arai R, Taniuchi T, Matsuura K, editors. Indo- also thank Julie Arruda, Scott Cramer, Judy Fen- Pacific Fish Biology: Proceedings of the Second International wick, and Anita Norton (WHOI) for their invalu- Conference on Indo-Pacific Fishes. Tokyo: Ichthyological Soci- ety of Japan. pp 570–589. able expertise and resourcefulness. They thank Chao NL. 1995. Sciaenidae. In: Fischer W, Krupp F, Schneider Drs. Mardi Hastings and David Zeddies (U. Mary- W, Sommer C, Carpenter KE, Niem VH, editors. Guı´aFAO land) for fruitful discussions and Drs. Timothy Tri- para la identificacı´on para los fines de la pesca. Pacifico Cen- cas (U. Hawaii), Allen Mensinger (U. Minnesota), tro-Oriental 3. Rome: Food and Agricultural Organization. pp and Richard Fay (Loyola U. Chicago) for comment- 1427–1528 Collette BB, Nauen CE. 1983. FAO Species Catalogue, Vol. 2: ing on earlier drafts of the manuscript. Under- Scombrids of the World. An Annotated and Illustrated Cata- graduates Erin Shearman, Ryan Walsh, Natasha logue of Tunas, Mackerels, Bonitos and Related Species Kelly, Nicole Cicchino, Sokun Ky, and Mary Turn- Known to Date. FAO Fish Synop No. 125:1–137. ipseed prepared and analyzed histological material Coombs S, Popper AN. 1979. Hearing differences among Hawai- ian squirrelfishes (family Holocentridae) related to differences used in this study. They thank Dr. Stanley Blum in the peripheral auditory system. J Comp Physiol 132:203– who provided Figure 3B and Melissa Tarby Wasko 207. who prepared Figure 1A. Portions of this work Coombs S, Montgomery JC. 1999. The enigmatic lateral-line were completed in partial fulfillment of require- system. In: Fay RR, Popper AN, editors. Comparative Hear- ments for the Master of Science degree at Villa- ing: Fish and Amphibians. New York: Springer-Verlag. pp 319–362. nova University to WLS. This work was supported Coombs S, Janssen J, Montgomery F. 1992. Functional and evo- by a Villanova Graduate Research Fellowship, a lutionary implications of peripheral diversity in lateral line Center for Environmental Research Graduate-Fac- systems. In: Webster D, Popper AN, Fay RR, editors. The Ev- ulty Fellowship (Columbia U.), an AMNH Lerner- olutionary Biology of Hearing. New York: Springer-Verlag. pp Gray Graduate Fellowship and an AMNH Kalb- 267–294. Dehadrai PV. 1959. On the swimbladder and its connection with fleisch Graduate Fellowship to WLS, an HHMI the internal ear in family Cichlidae. Proc Natl Inst Sci India Undergraduate Education grant to Villanova Uni- B 25:254–261. versity, and a Villanova Faculty Summer Research Dobbin CN. 1941. A comparative study of the gross anatomy of Fellowship to JFW. the air-bladders of ten families of fishes of New York and other eastern states. J Morphol 68:1–29. Fink SV, Fink WL. 1996. Interrelationships of ostariophysan fishes (Teleostei). In: Stiassny MLJ, Parenti LR, Johnson GD, editors. Interrelationships of Fishes. San Diego: Academic LITERATURE CITED Press. pp 209–249. Fletcher LB, Crawford JD. 2001. Acoustic detection by sound- Alexander RM. 1962. The structure of the Weberian apparatus producing fishes (Mormyridae): The role of gas-filled tympanic in the Cyprini. Proc Zool Soc Lond 139:451–473. bladders. J Exp Biol 204:175–183. Alexander RM. 1967. Functional design in fishes. London: Foote KG. 1980. Importance of the swimbladder in acoustic Hutchinson. scattering by fish: A comparison of gadoid and mackerel tar- Barimo JF, Fine ML. 1998. Relationship of swim-bladder shape get strengths. J Acoust Soc Am 67:2084–2089. to the directionality pattern of underwater sound in the oys- Foote KG. 1985. Rather-high-frequency sound scattering by ter toadfish. Can J Zool 76:134–143. swimbladdered fish. J Acoust Soc Am 78:688–700. Bauchot RA, Thomas A, Bauchot ML. 1989. The membranous Foote KG. 1988. Comparison of walleye pollack target strength labyrinth and its innervation in Chaetodon trifasciatus (Pis- estimates determined from in situ measurements and cal- ces, Teleostei, Chaetodontidae). Environ Biol Fishes 25:235– culations based on swimbladder form. J Acoust Soc Am 83: 242. 9–17. Bennett HS, Wyrick AD, Lee SW, McNeil JH Jr. 1976. Science Foote KG, Francis DTI. 2002. Acoustic scattering by swimblad- and art in preparing tissues embedded in plastic for light mi- dered fish: A review. Bioacoustics 12:263–265. croscopy, with special reference to glycol methacrylate, glass Gosline WA. 1966. The limits of the fish family Serranidae, with knives and simple stains. Stain Technol 51:71–97. notes on other lower percoids. Proc Calif Acad Sci 33:91–112. Best ACG, Gray JAB. 1980. Morphology of the utricular recess Green JM. 1971. Studies on the swim bladders of Eucinostomus in the sprat. J Mar Biol Assoc UK 60:703–715. gula and E. argenterus (Pisces: Gerreidae). Bull Mar Sci Blaxter, JHS. 1981. The swimbladder and hearing. In: Tavolga 21:567–590. WN, Popper AN, Fay RR, editors. Hearing and Sound Com- Greenwood PH. 1970. Skull and swimbladder connections in munication in Fishes. New York: Springer-Verlag. pp 61–70. the fishes of the family Megalopidae. Bull British Mus Nat Blaxter JHS, Tytler P. 1978. Physiology and function of the Hist (Zool) 14:121–135. swimbladder. Adv Comp Physiol Biochem 7:311–367. Gu¨ nther A. 1860. Catalogue of the Fishes in the British Blum SD. 1988. The osteology and phylogeny of the Chaetodon- Museum, Vol. 2: Squamipinnes, Cirrihitidae, Triglidae, Tra- tidae (Teleostei: Perciformes). Ph.D. Dissertation, University of chinidae, Sciaenidae, Polynemidae, Sphyraenidae, Trichiuri- Hawaii, Honolulu. Available at www.calacademy.org/research/ dae, Scombridae, Carangidae, Xiphiidae. London: British Mu- informatics/sblum. seum.

Journal of Morphology DOI 10.1002/jmor 1354 J.F. WEBB ET AL.

Hawkins AD. 1993. Underwater sound and fish behaviour. In: Pothoff T. 1984. Clearing and staining techniques. In: Moser Pitcher TJ, editor. Behaviour of Teleost Fishes, 2nd ed. Lon- HG, Richards WJ, Cohen D, Fahay MP, Kendall AW, Richard- don: Chapman and Hall. pp 129–170. son S, editors. Ontogeny and Systematics of Fishes. Law- Herre AW, Montalban HR. 1927. The Philippine butterflyfishes rence, KS: Allen Press. pp 35–37. and their allies. Philipp J Sci 34:1–113. Ramcharitar J, Higgs D, Popper AN. 2002. Sciaenid inner ears: Hourigan TF. 1989. Environmental determinants of butterfly- A study in diversity. Brain Behav Evol 58:152–162. fish social systems. Environ Biol Fish 25:61–78. Ramcharitar JU, Deng X, Ketten D, Popper AN. 2004. Form Johnson GD. 1980. The limits and relationships of the Lutjani- and function in the unique inner ear of a teleost: The dae and associated families. Bull Scripps Inst Oceanogr Univ silver perch (Bairdiella chrysoura). J Comp Neurol 475:531– Calif 24:1–114. 539. Kalmijn AJ. 1989. Functional evolution of lateral line and inner Reese ES. 1975. A comparative field study of the social behavior ear sensory systems. In: Coombs SP, Go¨rner P, Mu¨nz H, edi- and related ecology of reef fishes of the family Chaetodonti- tors. The Mechanosensory Lateral Line-Neurobiology and dae. Z Tierpsychol 37:37–61. Evolution. New York: Springer-Verlag. pp 187–215. Rosen DE, Greenwood PH. 1970. Origin of the Weberian Katayama M. 1959. Studies on the serranid fishes of Japan (I). ossicles and the relationships of the ostariophysan and gonor- Bull Fac Educ Yamaguichi Univ 8:103–180. ynchiform fishes. Amer Mus Novitates 2428:1–25. Kearney M, Maisano JA, Rowe T. 2005. Cranial anatomy of Rosenblatt RH, Bell MA. 1976. Osteology and relationships of theo extinct amphisbaenian Rhineura hatcherii (Squamata, the roosterfish. Nematistius pectoralis Gill. Nat Hist Mus Los Amphisbaenia) based on high-resolution X-ray computed to- Angeles County Contrib Sci 279:1–23. mography. J Morphol 264:1–33. Sand O, Hawkins AD. 1973. Acoustic properties of the cod Ketten DR, Skinner MW, Want G, Vannier MW, Gates GA, swimbladder. J Exp Biol 58:797–820. Neely JG. 1998. In vivo measures of cochlear length and Sasaki K. 1989. Phylogeny of the family Sciaenidae, with notes insertion depth of nucleus cochlear implant electrode arrays. on its zoogeography (Teleostei, Perciformes). Mem Fac Fish Ann Otol Rhinol LaryngolSuppl 175, 107(11, Part 2):1–16. Hokkaido Univ 36:1–137. Ladich F, Bass AH. 2003. Underwater sound generation and Schaefer KM, Oliver CW. 2000. Shape, volume, and resonance acoustic reception in fishes with some notes on frogs. In: frequency of the swimbladder of yellowfin tuna, Thunnus Collin SO, Marshall NJ, editors. Sensory Processing in albacares. US Fish Bull 98:364–374. Aquatic Environments. New York: Springer-Verlag. pp 173– Schaefer SA. 2003. Relationships of Lithogenes villosus Eigen- 193. mann, 1909 (Siluriformes, Loricariidae): Evidence from high- Ladich F, Popper AN. 2004. Parallel evolutoin in fish hearing resolution computed microtomography. Am Mus Novitates organs. In: Manley GA, Popper AN, Fay RR, editors. Evolu- 3401:55. tion of the Vertebrate Auditory System. New York: Springer- Schellart NAM, Popper AN. 1992. Functional aspects of the Verlag, pp 95–127. evolution of the auditory system of actinopterygian fish. Leis JM. 1994. Larvae, adults and relationships of the mono- In: Webster D, Popper AN, Fay RR, editors. The Evolution- typic perciform fish family Lactariidae. Rec Aust Mus 46:131– ary Biology of Hearing. New York: Springer-Verlag. pp 295– 143. 322. MacDonald CM. 1978. Morphological and biochemical system- Smith WL. 2000. Variation in the morphology of the laterophy- atics of Australian freshwater and estuarine percichthyid sic connection in butterflyfishes of the genus Chaetodon.MS fishes. Aust J Mar Freshwater Res 29:667–698. Thesis, Villanova University, Villanova, PA. MacLennan DN, Simmonds EJ. 1992. Fisheries Acoustics. Lon- Smith WL, Webb JF, Blum SD. 2003. The evolution of the later- don: Chapman and Hall. ophysic connection with a revised phylogeny and of McKay R. 1985. A revision of the fishes of the family Sillagini- butterflyfishes (Teleostei: Chaetodontidae). Cladistics 19:287– dae. Mem Queensland Mus 22:1–74. 306. Myrberg AA Jr. 1981. Sound communication and interception in Starnes WC. 1988. Revision, phylogeny and biogeographic com- fishes. In: Tavolga WN, Popper AN, Fay RR, editors. Hearing ments on the circumtropical marine percoid fish family Pria- and Sound Communication in Fishes. New York: Springer- canthidae. Bull Mar Sci 43:117–203. Verlag. pp 395–426. Steen JB. 1970. The swimbladder as a hydrostatic organ. In: Nelson EM. 1955. The morphology of the swim bladder and Hoar WS, Randall DJ, editors. Fish Physiology, Vol. 4. New auditory bulla in the Holocentridae. Fieldiana (Zool) 37:121– York: Academic Press. pp 413–443. 130. Summers AP, Ketcham RA, Rowe T. 2004. Structure and func- O’Connell CP. 1955. The gas bladder and its relation to the tion of the horn shark (Heterodontus francisci) cranium inner ear in Sardinops caerulea and Engraulis mordax.US through ontogeny: Development of a hard prey specialist. Fish Bull 56:505–533. J Morphol 260:1–12. Parker TH. 1882. On the connection of the air bladder and the Tavolga WN. 1974. Sensory parameters in communication auditory organ in the red cod (Lotella bacchus). Trans New among coral reef fishes. Mount Sinai J Med 41:324–340. Zealand Inst 15:234. Tricas TC. 1986. Life history, foraging ecology, and territorial Poggendorf D. 1952. The absolute threshold of hearing of the behavior of the Hawaiian butterflyfish, Chaetodon multi- bullhead (Amiurus nebulosus) and contributions to the physics cinctus. PhD Dissertation, University of Hawaii, Honolulu. of the Weberian apparatus of the Ostariophysi. (Adler HE, 248 p. Cappelli B, Transl.). Zeitschr Vergleigh Physiol 34:222–257. Tricas TC, Boyle KS. 2005. The evolution of pairing behavior, Popper AN. 1977. A scanning electron microscopic study of the sound production and hearing in chaetodontid butterflyfishes: sacculus and lagena in the ears of fifteen species of teleost Evidence from behavior and physiology. Brain Behav Evol fishes. J Morphol 153:397–417. 66:143. Popper AN, Fay RR. 1993. Sound detection and processing by Tricas TC, Boyle KS, Dale J. 2004. Sound production and social fish: Critical review and major research questions. Brain pairing in laterophysic butterflyfishes. Integr Comp Biol Behav Evol 41:14–38. 44:653. Popper AN, Fay RR. 1999. The auditory periphery in fishes. In: Wartzok D, Ketten DR. 1999. Marine mammal sensory systems. Fay RR, Popper AN, editors. Comparative Hearing: Fish and In: Reynolds JE, Rommel SA, editors. Biology of Marine Amphibians. New York: Springer-Verlag. pp 38–100. Mammals. Washington: Smithsonian Institution Press. pp Popper AN, Fay RR, Platt C, Sand O. 2003. Sound detection 117–175. mechanisms and capabilities of teleost fishes. In: Collin SO, Webb JF. 1998. Laterophysic connection: A unique link between Marshall NJ, editors. Sensory Processing in Aquatic Environ- the swimbladder and the lateral-line system in Chaetodon ments. New York: Springer-Verlag. pp 3–38. (Perciformes: Chaetodontidae). Copeia 1998:1032–1036.

Journal of Morphology DOI 10.1002/jmor CHAETODON LATEROPHYSIC CONNECTION 1355

Webb JF, Blum SD. 1990. A swimbladder-lateral-line connection Whitehead PJP, Blaxter JHS. 1989. Swimbladder form in clu- in the butterflyfish genus Chaetodon (Perciformes: Chaeto- peoid fishes. Zool J Linn Soc 97:299–372. dontidae). Am Zool 30:90A. Witmer LM. 2001. Nostril position in dinosaurs and other verte- Webb JF, Smith WL. 2000. The laterophysic connection in chae- brates and its significance for nasal function. Science 293: todontid butterflyfish: Morphological variation and specula- 850–853. tions on sensory function. Philos Trans R Soc Lond B Biol Sci Witmer LM, Chatterjee S, Franzosa J, Rowe T. 2003. Neuroan- 355:1125–1129. atomy of flying reptiles and implications for flight, posture Webb JF, Smith WL, Herman JL, Woods CF, Ketten DL. 2005. and behaviour. Nature 425:950–953. The laterophysic connection: Peripheral specialization for Yan HY, Curtsinger WS. 2000. The otic gasbladder as an ancil- reception of acoustic stimuli in chaetodontid butterflyfishes? lary auditory structure in a mormyrid fish. J Comp Physiol A Brain Behav Evol 66:137–144. 186:595–602.

Journal of Morphology DOI 10.1002/jmor